The main objective of the study is to enhance our knowledge of the factors affecting infant mortality, and to identify these factors that are most important and can be modified through policies and programs. Three aims of this project are to acquire additional data and integrate them into the existing data set, explore new analytical relationships and develop a working model of the infant mortality process, and introduce new methodological approaches for classifying and analyzing the data. The expanded data set will consist of linked birth, death, and multiple cause-of- death records for 1980 through 1984 birth cohorts in Florida. Available information on neonatal intensive care of some of the infants will also be integrated into the data files. Basic variables in the study are grouped under three headings -- Prior (mother's education, race/ethnicity, rural-urban residence, marital status); proximate (previous fetal loss, previous birth interval, mother's age, prenatal care, sex of baby, birth-weight, gestational age, hospital birth, birth order, plurality); and immediate (age at death, cause of death), Stress is placed on alternative cause-of- death measures, with emphasis on multiple causation. Event history (competing risks) analyses will be used along with other multivariate techniques for elaboration of a model specifying infant mortality as both a temporal and multidimensional process. The findings should permit policy and programmatic formulations that would help accelerate the decline of infant mortality.